Magnetic resonance imaging (MRI) systems are widely used in medical diagnose. For example, cardiac magnetic resonance imaging (CMR) techniques are widely used in a routine examination for the heart of a patient. As another example, real-time MRI techniques can continuously monitor an object (e.g., a patient) to be scanned by the MRI systems in real-time and continuously collect multiple consecutive images of the object during a time period. Currently, an MR cardiac cine technique may be used to reconstruct an image sequence of the heart by combining the cardiac magnetic resonance imaging (CMR) techniques and real-time MRI techniques. However, the MR cardiac cine techniques currently use an ECG-gating technique or a segmented acquisition technique to collecting MR signals, which needs a patient to hold a breath for a time period and can not be applied for a patient with arrhythmia. Thus, it is desirable to provide systems and methods for real-time MR image reconstruction with improved sampling.